Pixverse Pixverse V5.5 Text To Video

Pixverse Pixverse V5.5 Text To Video

Playground

PixVerse V5.5 transforms text prompts into realistic videos with smooth motion and natural detail in seconds—ideal for stories, ads, and social clips. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

PixVerse v5.5 — Text-to-Video

PixVerse v5.5 Text-to-Video turns a written scene description into a short animated clip. You control resolution (360p–1080p), duration (5s / 8s / 10s) and resolution_ratio (16:9,4:3,1:1,3:4,9:16), while the model handles camera motion, lighting and transitions for you.

✨ Highlights

• Multiple resolutions – 360p, 540p, 720p, 1080p for previews through to final export.
• Flexible aspect ratios – 16:9, 4:3, 1:1, 3:4, 9:16 to match feeds, stories and banners.
• Variable duration – 5, 8 or 10 seconds per clip.
• Prompt reasoning (thinking_type) – optional system-side enhancement that can refine and structure your prompt.
• Negative prompt support – steer the model away from artefacts such as “watermark”, “text”, “distortion”.
• Seed control – fix a seed for reproducible generations, or vary it for multiple takes.

🧩 Parameters

• prompt* (string) : Up to 2048 characters describing the scene, pacing and camera moves.

• resolution : One of 360p, 540p, 720p, 1080p.

• duration : 5, 8 or 10 seconds. (10s is not available for 1080p)

• resolution_ratio : 16:9, 4:3, 1:1, 3:4, 9:16.

• thinking_type :

  • Enabled: Turn on system-level reasoning/optimisation of your prompt.
  • Disabled: Use your prompt as-is.
  • Auto (default): Let the system decide whether to enable it.

• negative_prompt (optional) : Terms you don’t want to see in the video (e.g. watermark, text, logo, glitch).

• seed : Integer for reproducibility. Use a fixed value to re-run the same idea; change it to get new variations.

🚀 How to Use

1. Write the prompt

   • Describe key shots, mood and motion, e.g. “Anime rooftop at sunset, slow dolly-in toward the character, hair and clothes moving in the wind, cinematic lighting.”

2. Set resolution & resolution_ratio

   • Select resolution from 360p,540p,720p and 1080p.
   • 9:16 for TikTok, Reels.
   • 16:9 for Youtube videos
   • 1:1 for Instagram.
   • 4:3 for feed posts and thumbnails.

3. Choose duration

   • 5s for quick previews or punchy hooks.
   • 8s for more developed mini-scenes.

4. (Optional) Adjust thinking & negative_prompt

   • Set thinking_type=“enabled” if you want the system to help structure a complex prompt.
   • Add a short negative prompt to suppress text, artefacts or unwanted objects.

5. Set seed

   • Keep a seed fixed while you tweak the prompt.
   • Change it when you’re happy with the setup but want new takes.

6. Run and download

   • Generate the clip, review, then iterate on prompt, duration or aspect ratio as needed.

💡 Prompt & Quality Tips

• Write shot-by-shot for best results (wide → medium → close-up, etc.).
• Keep the number of major events small; let the model focus on a few strong beats.
• Higher resolutions (720p / 1080p) are better for export and editing; use 360p / 540p for fast generation.
• For vertical platforms, set **resolution_ratio = 9:16 or 3:4 ** from the start to avoid awkward cropping.

# Submit the task
curl --location --request POST "https://api.wavespeed.ai/api/v3/pixverse/pixverse-v5.5/text-to-video" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}" \
--data-raw '{
    "resolution": "540p",
    "duration": 5,
    "resolution_ratio": "1:1",
    "generate_audio_switch": false,
    "generate_multi_clip_switch": false,
    "thinking_type": "auto"
}'

# Get the result
curl --location --request GET "https://api.wavespeed.ai/api/v3/predictions/${requestId}/result" \
--header "Authorization: Bearer ${WAVESPEED_API_KEY}"